diff mbox series

[SRU,J:linux-bluefield,v3,02/11] net/sched: cls_api: Support hardware miss to tc action

Message ID 20230606230120.407292-3-witu@nvidia.com
State New
Headers show
Series Support hardware miss to tc action | expand

Commit Message

William Tu June 6, 2023, 11:01 p.m. UTC
From: Paul Blakey <paulb@nvidia.com>

BugLink: https://bugs.launchpad.net/bugs/2012571

For drivers to support partial offload of a filter's action list,
add support for action miss to specify an action instance to
continue from in sw.

CT action in particular can't be fully offloaded, as new connections
need to be handled in software. This imposes other limitations on
the actions that can be offloaded together with the CT action, such
as packet modifications.

Assign each action on a filter's action list a unique miss_cookie
which drivers can then use to fill action_miss part of the tc skb
extension. On getting back this miss_cookie, find the action
instance with relevant cookie and continue classifying from there.

Signed-off-by: Paul Blakey <paulb@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
(backported from commit 80cd22c35c9001fe72bf614d29439de41933deca)
[paulb: adjust for context differences and lack of tc_setup_action()]
Signed-off-by: Paul Blakey <paulb@nvidia.com>
Signed-off-by: William Tu <witu@nvidia.com>
---
 include/linux/skbuff.h     |   6 +-
 include/net/flow_offload.h |   1 +
 include/net/pkt_cls.h      |  30 ++++--
 include/net/sch_generic.h  |   2 +
 net/openvswitch/flow.c     |   3 +-
 net/sched/cls_api.c        | 215 +++++++++++++++++++++++++++++++++++--
 6 files changed, 235 insertions(+), 22 deletions(-)
diff mbox series

Patch

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 19e595cab23a..4f64a37d853e 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -284,12 +284,16 @@  struct nf_bridge_info {
  * and read by ovs to recirc_id.
  */
 struct tc_skb_ext {
-	__u32 chain;
+	union {
+		u64 act_miss_cookie;
+		__u32 chain;
+	};
 	__u16 mru;
 	__u16 zone;
 	u8 post_ct:1;
 	u8 post_ct_snat:1;
 	u8 post_ct_dnat:1;
+	u8 act_miss:1; /* Set if act_miss_cookie is used */
 };
 #endif
 
diff --git a/include/net/flow_offload.h b/include/net/flow_offload.h
index 5cc43cf3e075..ae618fda8922 100644
--- a/include/net/flow_offload.h
+++ b/include/net/flow_offload.h
@@ -199,6 +199,7 @@  struct flow_action_entry {
 	enum flow_action_id		id;
 	enum flow_action_hw_stats	hw_stats;
 	void				*act;
+	u64				miss_cookie;
 	action_destr			destructor;
 	void				*destructor_priv;
 	union {
diff --git a/include/net/pkt_cls.h b/include/net/pkt_cls.h
index 83a6d0792180..0644ba083115 100644
--- a/include/net/pkt_cls.h
+++ b/include/net/pkt_cls.h
@@ -59,6 +59,8 @@  int tcf_block_get_ext(struct tcf_block **p_block, struct Qdisc *q,
 void tcf_block_put(struct tcf_block *block);
 void tcf_block_put_ext(struct tcf_block *block, struct Qdisc *q,
 		       struct tcf_block_ext_info *ei);
+int tcf_exts_init_ex(struct tcf_exts *exts, struct net *net, int action,
+		     int police, struct tcf_proto *tp, u32 handle, bool used_action_miss);
 
 static inline bool tcf_block_shared(struct tcf_block *block)
 {
@@ -203,6 +205,7 @@  struct tcf_exts {
 	int nr_actions;
 	struct tc_action **actions;
 	struct net *net;
+	struct tcf_exts_miss_cookie_node *miss_cookie_node;
 #endif
 	/* Map to export classifier specific extension TLV types to the
 	 * generic extensions API. Unsupported extensions must be set to 0.
@@ -214,18 +217,11 @@  struct tcf_exts {
 static inline int tcf_exts_init(struct tcf_exts *exts, struct net *net,
 				int action, int police)
 {
-#ifdef CONFIG_NET_CLS_ACT
-	exts->type = 0;
-	exts->nr_actions = 0;
-	exts->net = net;
-	exts->actions = kcalloc(TCA_ACT_MAX_PRIO, sizeof(struct tc_action *),
-				GFP_KERNEL);
-	if (!exts->actions)
-		return -ENOMEM;
+#ifdef CONFIG_NET_CLS
+	return tcf_exts_init_ex(exts, net, action, police, NULL, 0, false);
+#else
+	return -EOPNOTSUPP;
 #endif
-	exts->action = action;
-	exts->police = police;
-	return 0;
 }
 
 /* Return false if the netns is being destroyed in cleanup_net(). Callers
@@ -317,6 +313,18 @@  tcf_exts_exec(struct sk_buff *skb, struct tcf_exts *exts,
 	return TC_ACT_OK;
 }
 
+static inline int
+tcf_exts_exec_ex(struct sk_buff *skb, struct tcf_exts *exts, int act_index,
+		 struct tcf_result *res)
+{
+#ifdef CONFIG_NET_CLS_ACT
+	return tcf_action_exec(skb, exts->actions + act_index,
+			       exts->nr_actions - act_index, res);
+#else
+	return TC_ACT_OK;
+#endif
+}
+
 int tcf_exts_validate(struct net *net, struct tcf_proto *tp,
 		      struct nlattr **tb, struct nlattr *rate_tlv,
 		      struct tcf_exts *exts, u32 flags,
diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h
index 6906da5c733e..5bf56737a664 100644
--- a/include/net/sch_generic.h
+++ b/include/net/sch_generic.h
@@ -367,6 +367,8 @@  struct tcf_proto_ops {
 						struct nlattr **tca,
 						struct netlink_ext_ack *extack);
 	void			(*tmplt_destroy)(void *tmplt_priv);
+	struct tcf_exts *	(*get_exts)(const struct tcf_proto *tp,
+					    u32 handle);
 
 	/* rtnetlink specific */
 	int			(*dump)(struct net*, struct tcf_proto*, void *,
diff --git a/net/openvswitch/flow.c b/net/openvswitch/flow.c
index 1b81d71bac3c..e6b8966ddb54 100644
--- a/net/openvswitch/flow.c
+++ b/net/openvswitch/flow.c
@@ -897,7 +897,8 @@  int ovs_flow_key_extract(const struct ip_tunnel_info *tun_info,
 #if IS_ENABLED(CONFIG_NET_TC_SKB_EXT)
 	if (static_branch_unlikely(&tc_recirc_sharing_support)) {
 		tc_ext = skb_ext_find(skb, TC_SKB_EXT);
-		key->recirc_id = tc_ext ? tc_ext->chain : 0;
+		key->recirc_id = tc_ext && !tc_ext->act_miss ?
+				 tc_ext->chain : 0;
 		OVS_CB(skb)->mru = tc_ext ? tc_ext->mru : 0;
 		post_ct = tc_ext ? tc_ext->post_ct : false;
 		post_ct_snat = post_ct ? tc_ext->post_ct_snat : false;
diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index 9574bf9037b9..5911e83388e7 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -22,6 +22,7 @@ 
 #include <linux/idr.h>
 #include <linux/jhash.h>
 #include <linux/rculist.h>
+#include <linux/rhashtable.h>
 #include <net/net_namespace.h>
 #include <net/sock.h>
 #include <net/netlink.h>
@@ -49,6 +50,109 @@  static LIST_HEAD(tcf_proto_base);
 /* Protects list of registered TC modules. It is pure SMP lock. */
 static DEFINE_RWLOCK(cls_mod_lock);
 
+static struct xarray tcf_exts_miss_cookies_xa;
+struct tcf_exts_miss_cookie_node {
+	const struct tcf_chain *chain;
+	const struct tcf_proto *tp;
+	const struct tcf_exts *exts;
+	u32 chain_index;
+	u32 tp_prio;
+	u32 handle;
+	u32 miss_cookie_base;
+	struct rcu_head rcu;
+};
+
+/* Each tc action entry cookie will be comprised of 32bit miss_cookie_base +
+ * action index in the exts tc actions array.
+ */
+union tcf_exts_miss_cookie {
+	struct {
+		u32 miss_cookie_base;
+		u32 act_index;
+	};
+	u64 miss_cookie;
+};
+
+#if IS_ENABLED(CONFIG_NET_TC_SKB_EXT)
+static int
+tcf_exts_miss_cookie_base_alloc(struct tcf_exts *exts, struct tcf_proto *tp,
+				u32 handle)
+{
+	struct tcf_exts_miss_cookie_node *n;
+	static u32 next;
+	int err;
+
+	if (WARN_ON(!handle || !tp->ops->get_exts))
+		return -EINVAL;
+
+	n = kzalloc(sizeof(*n), GFP_KERNEL);
+	if (!n)
+		return -ENOMEM;
+
+	n->chain_index = tp->chain->index;
+	n->chain = tp->chain;
+	n->tp_prio = tp->prio;
+	n->tp = tp;
+	n->exts = exts;
+	n->handle = handle;
+
+	err = xa_alloc_cyclic(&tcf_exts_miss_cookies_xa, &n->miss_cookie_base,
+			      n, xa_limit_32b, &next, GFP_KERNEL);
+	if (err)
+		goto err_xa_alloc;
+
+	exts->miss_cookie_node = n;
+	return 0;
+
+err_xa_alloc:
+	kfree(n);
+	return err;
+}
+
+static void tcf_exts_miss_cookie_base_destroy(struct tcf_exts *exts)
+{
+	struct tcf_exts_miss_cookie_node *n;
+
+	if (!exts->miss_cookie_node)
+		return;
+
+	n = exts->miss_cookie_node;
+	xa_erase(&tcf_exts_miss_cookies_xa, n->miss_cookie_base);
+	kfree_rcu(n, rcu);
+}
+
+static struct tcf_exts_miss_cookie_node *
+tcf_exts_miss_cookie_lookup(u64 miss_cookie, int *act_index)
+{
+	union tcf_exts_miss_cookie mc = { .miss_cookie = miss_cookie, };
+
+	*act_index = mc.act_index;
+	return xa_load(&tcf_exts_miss_cookies_xa, mc.miss_cookie_base);
+}
+#else /* IS_ENABLED(CONFIG_NET_TC_SKB_EXT) */
+static int
+tcf_exts_miss_cookie_base_alloc(struct tcf_exts *exts, struct tcf_proto *tp,
+				u32 handle)
+{
+	return 0;
+}
+
+static void tcf_exts_miss_cookie_base_destroy(struct tcf_exts *exts)
+{
+}
+#endif /* IS_ENABLED(CONFIG_NET_TC_SKB_EXT) */
+
+static u64 tcf_exts_miss_cookie_get(u32 miss_cookie_base, int act_index)
+{
+	union tcf_exts_miss_cookie mc = { .act_index = act_index, };
+
+	if (!miss_cookie_base)
+		return 0;
+
+	mc.miss_cookie_base = miss_cookie_base;
+	return mc.miss_cookie;
+}
+
 static u32 destroy_obj_hashfn(const struct tcf_proto *tp)
 {
 	return jhash_3words(tp->chain->index, tp->prio,
@@ -1529,6 +1633,8 @@  static inline int __tcf_classify(struct sk_buff *skb,
 				 const struct tcf_proto *orig_tp,
 				 struct tcf_result *res,
 				 bool compat_mode,
+				 struct tcf_exts_miss_cookie_node *n,
+				 int act_index,
 				 u32 *last_executed_chain)
 {
 #ifdef CONFIG_NET_CLS_ACT
@@ -1540,13 +1646,36 @@  static inline int __tcf_classify(struct sk_buff *skb,
 #endif
 	for (; tp; tp = rcu_dereference_bh(tp->next)) {
 		__be16 protocol = skb_protocol(skb, false);
-		int err;
+		int err = 0;
 
-		if (tp->protocol != protocol &&
-		    tp->protocol != htons(ETH_P_ALL))
-			continue;
+		if (n) {
+			struct tcf_exts *exts;
+
+			if (n->tp_prio != tp->prio)
+				continue;
+
+			/* We re-lookup the tp and chain based on index instead
+			 * of having hard refs and locks to them, so do a sanity
+			 * check if any of tp,chain,exts was replaced by the
+			 * time we got here with a cookie from hardware.
+			 */
+			if (unlikely(n->tp != tp || n->tp->chain != n->chain ||
+				     !tp->ops->get_exts))
+				return TC_ACT_SHOT;
+
+			exts = tp->ops->get_exts(tp, n->handle);
+			if (unlikely(!exts || n->exts != exts))
+				return TC_ACT_SHOT;
+
+			n = NULL;
+			err = tcf_exts_exec_ex(skb, exts, act_index, res);
+		} else {
+			if (tp->protocol != protocol &&
+			    tp->protocol != htons(ETH_P_ALL))
+				continue;
 
-		err = tp->classify(skb, tp, res);
+			err = tp->classify(skb, tp, res);
+		}
 #ifdef CONFIG_NET_CLS_ACT
 		if (unlikely(err == TC_ACT_RECLASSIFY && !compat_mode)) {
 			first_tp = orig_tp;
@@ -1562,6 +1691,9 @@  static inline int __tcf_classify(struct sk_buff *skb,
 			return err;
 	}
 
+	if (unlikely(n))
+		return TC_ACT_SHOT;
+
 	return TC_ACT_UNSPEC; /* signal: continue lookup */
 #ifdef CONFIG_NET_CLS_ACT
 reset:
@@ -1586,21 +1718,35 @@  int tcf_classify(struct sk_buff *skb,
 #if !IS_ENABLED(CONFIG_NET_TC_SKB_EXT)
 	u32 last_executed_chain = 0;
 
-	return __tcf_classify(skb, tp, tp, res, compat_mode,
+	return __tcf_classify(skb, tp, tp, res, compat_mode, NULL, 0,
 			      &last_executed_chain);
 #else
 	u32 last_executed_chain = tp ? tp->chain->index : 0;
+	struct tcf_exts_miss_cookie_node *n = NULL;
 	const struct tcf_proto *orig_tp = tp;
 	struct tc_skb_ext *ext;
+	int act_index = 0;
 	int ret;
 
 	if (block) {
 		ext = skb_ext_find(skb, TC_SKB_EXT);
 
-		if (ext && ext->chain) {
+		if (ext && (ext->chain || ext->act_miss)) {
 			struct tcf_chain *fchain;
+			u32 chain;
+
+			if (ext->act_miss) {
+				n = tcf_exts_miss_cookie_lookup(ext->act_miss_cookie,
+								&act_index);
+				if (!n)
+					return TC_ACT_SHOT;
 
-			fchain = tcf_chain_lookup_rcu(block, ext->chain);
+				chain = n->chain_index;
+			} else {
+				chain = ext->chain;
+			}
+
+			fchain = tcf_chain_lookup_rcu(block, chain);
 			if (!fchain)
 				return TC_ACT_SHOT;
 
@@ -1612,7 +1758,7 @@  int tcf_classify(struct sk_buff *skb,
 		}
 	}
 
-	ret = __tcf_classify(skb, tp, orig_tp, res, compat_mode,
+	ret = __tcf_classify(skb, tp, orig_tp, res, compat_mode, n, act_index,
 			     &last_executed_chain);
 
 	/* If we missed on some chain */
@@ -3022,9 +3168,48 @@  static int tc_dump_chain(struct sk_buff *skb, struct netlink_callback *cb)
 	return skb->len;
 }
 
+int tcf_exts_init_ex(struct tcf_exts *exts, struct net *net, int action,
+		     int police, struct tcf_proto *tp, u32 handle,
+		     bool use_action_miss)
+{
+	int err = 0;
+
+#ifdef CONFIG_NET_CLS_ACT
+	exts->type = 0;
+	exts->nr_actions = 0;
+	/* Note: we do not own yet a reference on net.
+	 * This reference might be taken later from tcf_exts_get_net().
+	 */
+	exts->net = net;
+	exts->actions = kcalloc(TCA_ACT_MAX_PRIO, sizeof(struct tc_action *),
+				GFP_KERNEL);
+	if (!exts->actions)
+		return -ENOMEM;
+#endif
+
+	exts->action = action;
+	exts->police = police;
+
+	if (!use_action_miss)
+		return 0;
+
+	err = tcf_exts_miss_cookie_base_alloc(exts, tp, handle);
+	if (err)
+		goto err_miss_alloc;
+
+	return 0;
+
+err_miss_alloc:
+	tcf_exts_destroy(exts);
+	return err;
+}
+EXPORT_SYMBOL(tcf_exts_init_ex);
+
 void tcf_exts_destroy(struct tcf_exts *exts)
 {
 #ifdef CONFIG_NET_CLS_ACT
+	tcf_exts_miss_cookie_base_destroy(exts);
+
 	if (exts->actions) {
 		tcf_action_destroy(exts->actions, TCA_ACT_UNBIND);
 		kfree(exts->actions);
@@ -3557,6 +3742,7 @@  int tc_setup_flow_action(struct flow_action *flow_action,
 			 const struct tcf_exts *exts)
 {
 	struct tc_action *act;
+	u32 miss_cookie_base;
 	int i, j, k, err = 0;
 
 	BUILD_BUG_ON(TCA_ACT_HW_STATS_ANY != FLOW_ACTION_HW_STATS_ANY);
@@ -3566,6 +3752,8 @@  int tc_setup_flow_action(struct flow_action *flow_action,
 	if (!exts)
 		return 0;
 
+	miss_cookie_base = exts->miss_cookie_node ?
+			   exts->miss_cookie_node->miss_cookie_base : 0;
 	j = 0;
 	tcf_exts_for_each_action(i, act, exts) {
 		struct flow_action_entry *entry;
@@ -3573,12 +3761,16 @@  int tc_setup_flow_action(struct flow_action *flow_action,
 		entry = &flow_action->entries[j];
 		entry->act = act;
 		spin_lock_bh(&act->tcfa_lock);
+
 		err = tcf_act_get_cookie(entry, act);
 		if (err)
 			goto err_out_locked;
 
 		entry->hw_stats = tc_act_hw_stats(act->hw_stats);
 
+		if (!is_tcf_pedit(act))
+			entry->miss_cookie = tcf_exts_miss_cookie_get(miss_cookie_base, i);
+
 		if (is_tcf_gact_ok(act)) {
 			entry->id = FLOW_ACTION_ACCEPT;
 		} else if (is_tcf_gact_shot(act)) {
@@ -3646,6 +3838,9 @@  int tc_setup_flow_action(struct flow_action *flow_action,
 				entry->mangle.val = tcf_pedit_val(act, k);
 				entry->mangle.offset = tcf_pedit_offset(act, k);
 				entry->hw_stats = tc_act_hw_stats(act->hw_stats);
+				entry->miss_cookie =
+					tcf_exts_miss_cookie_get(miss_cookie_base, i);
+
 				entry = &flow_action->entries[++j];
 			}
 		} else if (is_tcf_csum(act)) {
@@ -3902,6 +4097,8 @@  static int __init tc_filter_init(void)
 	if (err)
 		goto err_register_pernet_subsys;
 
+	xa_init_flags(&tcf_exts_miss_cookies_xa, XA_FLAGS_ALLOC1);
+
 	rtnl_register(PF_UNSPEC, RTM_NEWTFILTER, tc_new_tfilter, NULL,
 		      RTNL_FLAG_DOIT_UNLOCKED);
 	rtnl_register(PF_UNSPEC, RTM_DELTFILTER, tc_del_tfilter, NULL,